If you can point to a better model of a more successful kernel that would be interesting to read about.

FreeBSD? But there is nothing fundamentally different in open-source project management between a kernel and any other large open-source project. The linux code base is the largest, but not by a large margin. Chrome, GCC, OpenOffice, Android (excluding the linux part obvs.), and the various BSDs are all comparable in scope, complexity, lines of code, and number of contributors. Only linux is (in)famous for having a toxic and unproductive culture.

Odd. Linux is by far the most/largest/longest successful project. Or do you have a counterexample?

I assume we're only comparing to operating systems, as I would say that e.g. Chrome as an open-source project is just as impactful if not more so. But generally speaking the Linux support experience is decidedly worse than *BSD or any proprietary OS. It is much harder to get patches upstreamed, and many hardware have errata that never get fixed. (If you haven't experienced this, it might be because companies like canonical and red hat maintain their own patches to support their customers.)

> But generally speaking the Linux support experience is decidedly worse than *BSD or any proprietary OS.

Going to have to disagree on that one.

Good for who? External parties that want to force their vision of what the project should be?